Dynamic parameter architecture for QP smoother

Abstract

According to an exemplary method, a smoothing module can be used in an ADV to iteratively perform a smoothing operation on a raw reference segment using an ordered list of sets of smoothing parameters, starting from the set of strictest parameters, until the smoothing operation is successful. The method includes the operations of generating multiple sets of smoothing parameters, including a first set of smoothing parameters, a second set of smoothing parameters, and at least one set of smoothing parameters interpolated in between; for each set of smoothing parameters, performing a quadratic programming (QP) smoothing operation on the raw reference line segment until the QP smoothing operation is successful; and controlling the ADV according to a smoothed reference line segment generated by the successful QP smoothing operation.

What is claimed is: 
     
       1. A computer-implemented method of operating an autonomous driving vehicle (ADV), the method comprising:
 generating a plurality of sets of smoothing parameters, including a first set of smoothing parameters, a second set of smoothing parameters that are looser than the first set, and at least one set of smoothing parameters interpolated between the first set and the second set of smoothing parameters; 
 for each of the plurality of sets of smoothing parameters, performing a quadratic programming (QP) smoothing operation on a segment of a raw reference line until the QP smoothing operation satisfies a predetermined condition; and 
 planning a trajectory to control the ADV according to a smoothed reference line segment generated by the QP smoothing operation. 
 
     
     
       2. The method of  claim 1 , wherein each of the plurality of sets of smoothing parameters includes a longitudinal freedom, a lateral freedom, a degree of polynomial functions, and a number of control points selected on the segment of the raw reference line. 
     
     
       3. The method of  claim 2 , wherein the longitudinal freedom and the lateral freedom in the first set of smoothing parameters are respectively smaller than the longitudinal freedom and the lateral freedom in the second set of smoothing parameters. 
     
     
       4. The method of  claim 3 , wherein the plurality of sets of smoothing parameters have a same degree of polynomial functions and a same number of control points selected on the segment of the raw reference line, but different longitudinal freedoms and different lateral freedoms. 
     
     
       5. The method of  claim 3 , wherein the plurality of sets of smoothing parameters are ordered ascendingly, starting with the first set of smoothing parameters. 
     
     
       6. The method of  claim 3 , wherein a ratio between the longitudinal freedom and the lateral freedom in each set of smoothing parameters is the same. 
     
     
       7. The method of  claim 1 , wherein performing the QP smoothing operation for each of the plurality of sets of smoothing parameters further comprises:
 determining a boundary box centered around each of a number of control points on the first segment of the raw reference line, wherein the boundary box has a longitudinal freedom of the set of smoothing parameters as a length and a lateral freedom of the set of smoothing parameters as a width; 
 selecting a plurality of two-dimensional polynomials each representing a segment of an optimal reference line between adjacent control points; and 
 defining a set of constraints to the two-dimensional polynomials to enable the plurality of two-dimensional polynomials to pass through the boundary box. 
 
     
     
       8. A non-transitory machine-readable medium having instructions stored therein for storing point cloud data in an autonomous driving vehicle (ADV), the instructions, when executed by a processor, cause the processor to perform operations, the operations comprising:
 generating a plurality of sets of smoothing parameters, including a first set of smoothing parameters, a second set of smoothing parameters that are looser than the first set, and at least one set of smoothing parameters interpolated between the first set and the second set of smoothing parameters; 
 for each of the plurality of sets of smoothing parameters, performing a quadratic programming (QP) smoothing operation on a segment of a raw reference line until the QP smoothing operation satisfies a predetermined condition; and 
 planning a trajectory to control the ADV according to a smoothed reference line segment generated by the QP smoothing operation. 
 
     
     
       9. The non-transitory machine-readable medium of  claim 8 , wherein each of the plurality of sets of smoothing parameters includes a longitudinal freedom, a lateral freedom, a degree of polynomial functions, and a number of control points selected on the segment of the raw reference line. 
     
     
       10. The non-transitory machine-readable medium of  claim 9 , wherein the longitudinal freedom and the lateral freedom in the first set of smoothing parameters are respectively smaller than the longitudinal freedom and the lateral freedom in the second set of smoothing parameters. 
     
     
       11. The non-transitory machine-readable medium of  claim 10 , wherein the plurality of sets of smoothing parameters have a same degree of polynomial functions and a same number of control points selected on the segment of the raw reference line, but different longitudinal freedoms and different lateral freedoms. 
     
     
       12. The non-transitory machine-readable medium of  claim 10 , wherein the plurality of sets of smoothing parameters are ordered ascendingly, starting with the first set of smoothing parameters. 
     
     
       13. The non-transitory machine-readable medium of  claim 10 , wherein a ratio between the longitudinal freedom and the lateral freedom in each set of smoothing parameters is the same. 
     
     
       14. The non-transitory machine-readable medium of  claim 8 , wherein performing the QP smoothing operation for each of the plurality of sets of smoothing parameters further comprises:
 determining a boundary box centered around each of a number of control points on the first segment of the raw reference line, wherein the boundary box has a longitudinal freedom of the set of smoothing parameters as a length and a lateral freedom of the set of smoothing parameters as a width; 
 selecting a plurality of two-dimensional polynomials each representing a segment of an optimal reference line between adjacent control points; and 
 defining a set of constraints to the two-dimensional polynomials to enable the plurality of two-dimensional polynomials to pass through the boundary box. 
 
     
     
       15. A data processing system, comprising:
 a processor; and 
 a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations, the operations including
 generating a plurality of sets of smoothing parameters, including a first set of smoothing parameters, a second set of smoothing parameters that are looser than the first set, and at least one set of smoothing parameters interpolated between the first set and the second set of smoothing parameters, 
 for each of the plurality of sets of smoothing parameters, performing a quadratic programming (QP) smoothing operation on a segment of a raw reference line until the QP smoothing operation satisfies a predetermined condition, and 
 planning a trajectory to control an autonomous driving vehicle (ADV) according to a smoothed reference line segment generated by the QP smoothing operation. 
 
 
     
     
       16. The system of  claim 15 , wherein each of the plurality of sets of smoothing parameters includes a longitudinal freedom, a lateral freedom, a degree of polynomial functions, and a number of control points selected on the segment of the raw reference line. 
     
     
       17. The system of  claim 16 , wherein the longitudinal freedom and the lateral freedom in the first set of smoothing parameters are respectively smaller than the longitudinal freedom and the lateral freedom in the second set of smoothing parameters. 
     
     
       18. The system of  claim 17 , wherein the plurality of sets of smoothing parameters have a same degree of polynomial functions and a same number of control points selected on the segment of the raw reference line, but different longitudinal freedoms and different lateral freedoms. 
     
     
       19. The system of  claim 17 , wherein the plurality of sets of smoothing parameters are ordered ascendingly, starting with the first set of smoothing parameters. 
     
     
       20. The system of  claim 17 , wherein a ratio between the longitudinal freedom and the lateral freedom in each set of smoothing parameters is the same.